BIOLOGICAL OVERVIEW

Genetic analysis in Drosophila has been used to study the process of C-terminal peptide alpha-amidation. This process is a late event in the biosynthesis of secretory peptides and is likely, in many instances, to be the rate-limiting step. In insects, more than 90% of known or predicted neuropeptides are amidated. Peptidylglycine alpha-hydroxylating monooxygenase (Phm) carries out the penultimate step in alpha-amidation, hydroxylating intermediates from prohormone precursor cleavage products that terminate in glycine residues (Jiang, 2000).

Phm mutants lack Phm protein and enzyme activity; most null animals die as late embryos with few morphological defects. Natural and synthetic Phm hypomorphs reveal phenotypes that resemble those of animals with mutations in genes of the ecdysone-inducible regulatory circuit. Animals bearing a strong hypomorphic allele contain no detectable Phm enzymatic activity or protein; ~50% hatch and initially display normal behavior, then die as young larvae, often while attempting to molt. Phm mutants are rescued with daily induction of a Phm transgene and complete rescue is seen with induction limited to the first 4 days after egg-laying. The rescued mutant adults produce progeny that survive to various stages up through metamorphosis (synthetic hypomorphs) and display prepupal and pupal phenotypes resembling those of ecdysone-response gene mutations. Examination of neuropeptide biosynthesis in Phm mutants reveals specific disruptions: amidated peptides are largely absent in strong hypomorphs, but peptide precursors, a nonamidated neuropeptide, nonpeptide transmitters, and other peptide biosynthetic enzymes are readily detected. Mutant adults that are produced by a minimal rescue schedule have lowered Phm enzyme levels and reproducibly altered patterns of amidated neuropeptides in the CNS. These deficits are partially reversed within 24 h by a single Phm induction in the adult stage. These genetic results support the hypothesis that secretory peptide signaling is critical for transitions between developmental stages, without strongly affecting morphogenetic events within a stage. Further, they show that Phm is required for peptide alpha-amidating activity throughout the life of Drosophila. Finally, they define novel methods to study neural and endocrine peptide biosynthesis and functions in vivo (Jiang, 2000).

C-terminal alpha-amidation results from the sequential actions of two enzymes: Phm and Pal (Peptidyl-alpha-hydroxyglycine-alpha-amidating lyase). While Phm creates hydroxylated intermediates from prohormone precursor cleavage products, Pal cleaves the intermediates to produce the final amidated peptides and glyoxylate. In vertebrates, the two enzymes occupy adjacent domains of a bifunctional protein called PAM (Eipper, 1992); in Drosophila, the Phm and Pal enzyme activities are both present, yet they are physically and genetically distinct (Kolhekar, 1997). The Drosophila genome sequencing project predicts one unlinked Phm gene (CG3832 at 60B1-2) and two unlinked Pal genes (CG12130 at 46C6-7 and CG5472 at 59F4-6). A homozygous lethal transposon insertion (P[07623]) lies within the coding region of Phm and reduces Phm enzyme levels (as measured in heterozygous adults) without also reducing Pal levels (Kolhekar, 1997 and Jiang, 2000).

The nervous systems of Phm^P[07623] homozygous, hemizygous, and heterozygous first-instar larvae were examined to determine whether Phm mutants display alterations in neuropeptide processing. An immunological approach was used to distinguish between amidated and nonamidated products derived from the pro-dFMRF precursor. Antiserum PT2 was generated against the tetrapeptide FMRFamide and reveals a pattern of staining that is greater than that displayed by products of the dFMRFamide gene. That pattern likely includes products of related neuropeptide genes (other peptides that share a common '-RFamide' C terminus). The second antiserum was directed against the final 19 amino acids of the pro-dFMRF prohormone, which is not amidated. The pro-dFMRF antiserum produces a pattern highly similar to that displayed by FMRFamide-related mRNA and by large FMRFamide-related-lacZ reporter transgenes. The PT2 antiserum stained a robust pattern of ~26 neurons in the larval brain and ventral ganglion of first-instar Phm^P[07623] heterozygotes, but revealed virtually nothing in the CNS of Phm^P[07623] homozygotes or hemizygotes. On rare occasions, small spots of immunoreactivity were observed in the CNSs of homozygotes in locations normally occupied by prominent FMRFamide-positive neurons. These 'spots' may be explained by any of several possibilities: (1) an alternate (low level or highly inefficient) source of Phm-like enzyme activity, (2) the action of maternally derived Phm mRNA or Phm enzyme activity, or (3) a lack of specificity by the anti-FMRFamide antiserum such that it weakly detects nonamidated peptide forms. In contrast, no differences were seen in the pattern or intensity of the ~14 neurons immunostained with the pro-dFMRF antiserum between heterozygotes and mutant Phm^P[07623] animals. These results are consistent with the hypothesis that Phm mutant animals have defects in a posttranslational step(s) in neuropeptide biosynthesis (Jiang, 2000).

To extend the phenotypic analysis to other amidated neuropeptides, antibodies were used to Aplysia peptide myomodulin (MM); there are ~34 MM-immunoreactive neurons in the larval CNS; antibody to Leucophaea leukokinin I (LKI: induces Leucophaea maderae hindgut contraction) stains a different set of ~16 neurons. In neither case have the putative, homologous peptides of Drosophila been isolated. Likewise, antibodies to nonamidated versions of myomodulin or leukokinin I are not available. It was found that these two antisera produced strong staining reactions in Phm^P[07623] heterozygous animals and weak or negligible staining reactions in Phm^P[07623] homozygous or hemizygous animals (Jiang, 2000).

Weak signals (in positions of neurons that are normally strongly stained) were seen in fewer than half the specimens and in only certain positions; zero staining was seen in most specimens. The anti-LKI staining pattern is not as intense as the MM pattern in wild-type animals; zero staining was seen in homozygous and in hemizygous Phm^P[07623] mutant animals. Together these immunochemical results support the hypothesis that Phm is required for the normal production of amidated neuropeptides in Drosophila larvae. Additional observations indicate that mutation of the Phm locus does not overtly disrupt aminergic transmitter systems, or those few peptidergic transmitter systems that normally lack amidation, or other neuropeptide processing enzymes (Jiang, 2000).

In Drosophila, mutations that specifically affect genes encoding secretory peptides are rare. This scarcity is mainly due to the difficulty in predicting accurate phenotypes for gene products that encode multiple signals whose physiological actions display both intra- and intergenic redundancy. Genetic methods to ablate specific secretory cells represent an alternative approach in vivo to studying secretory peptide functions. An alternative genetic approach is the study of secretory peptide biosynthesis. Kolhekar (1997a) and Siekhaus (1999) analyzed two separate genes encoding different neuropeptide-processing enzymes, Phm and PC2 (amontillado, the Drosophila homolog of the neuropeptide precursor processing protease), respectively. The genetics of neuropeptide processing offers a broad-based approach to examining secretory peptide functions. Such phenotypes provide insight into functions that secretory molecules may perform as a class (Jiang, 2000).

The P(28) chromosome contains 1309 bp deleted to one side of the inserted P[07623] transposon. The deletion removes critical Phm sequences, but also truncates an overlapping gene, CG17263. This other gene encodes a LIM-only protein, and the P(28) chromosome removes two of its three LIM domains. Because of this, the P(28) chromosome should also be considered an allele of CG17263. It is suspected that some part of the P(28) mutant phenotype may derive from absence of CG17263 function, but the contribution of CG17263 to the phenotype cannot be defined at present. The analysis of P(28) mutant phenotypes has been included in this analysis of Phm based on a strict reliance upon rescue with wild-type Phm sequences (Jiang, 2000).

The reversal of mutant phenotypes following heat shock-Phm induction permitted the inference that the absence of Phm is responsible for particular deficits. Specifically, full rescue of embryonic, larval, pupal, and adult lethality is possible. In addition, hs-Phm is capable of reversing the aberrant processing of secretory peptides (amidation of FMRF peptides). Therefore despite an ambiguity derived from the closely overlapped nature of the Phm and CG17263 genes, it is felt that specific interpretations concerning Phm functions are conservative and appropriate (Jiang, 2000).

Phm mutant animals die early in development, either as late embryos or as young larvae. The earliest lethal phase is seen with the Phm^P(28) allele, which is a null by several measures. Most mutant animals (as homozygotes and hemizygotes) reach late stages of embryogenesis with relatively normal morphological appearance. The CNS is slightly smaller than normal, but its organization and complexity appear normal. Animals homozygous for the Phm^P[07623] allele survive to later larval stages, as compared to Phm^P(28) mutants. While they contained no detectable Phm protein, they display potentially trace amounts of Phm enzyme activity. It is proposed that Phm^P[07623] animals, despite the insertion of a large transposon in the Phm locus, contain low levels of zygotic Phm enzyme due to compensatory transcriptional and/or posttranscriptional mechanisms. In contrast, Phm^P(28) animals, lacking about half of the Phm coding sequence, contain no zygotic enzyme activity (Jiang, 2000).

The relatively normal growth of mutant animals indicates that zygotic Phm expression is largely dispensable to complete the generation and morphogenesis of embryonic tissues. The contribution of maternally derived Phm to early morphogenesis is not yet known. While some larval tissues in the most severe Phm mutant sometimes appear smaller (e.g., the brain), a quantitative analysis is required to determine if and when these differences are significant. These results suggest a general view that, in insects, large-scale alterations of secretory peptide biosynthesis do not produce large-scale morphogenetic defects (Jiang, 2000).

Many Phm^P[07623] homozygotes die as late stage embryos or in the midst of larval molts. In general, these results indicate that Phm mutant animals have difficulty at or near times of developmental transitions: embryonic hatching and/or larval molting. A similar conclusion was reached by Siekhaus (1999) in an analysis of the PC2/amontillado gene. PC2/amon encodes a potential prohormone convertase with many similarities to the mammalian enzyme PC2, which is known to be important for processing of neural and endocrine peptides. A disruption of molting processes in Phm mutant animals is consistent with a large body of evidence relating secretory peptides to the orchestration of molting events (reviewed by Henrich, 1999). In particular, ecdysis (which is a late event in the molting process) is coordinated and modified by cascades of homones. These hormone cascades include several amidated secretory peptides made in the CNS or peripheral endocrine centers. The GAL4 3' UAS system will be useful to create tissue-specific Phm mosaics and so the question becomes which tissues must produce amidated peptides to permit normal embryonic hatching and larval molting (Jiang, 2000)?

First generation rescued mosaic animals are normal in many respects, but also show stereotyped behavioral abnormalities. Rescued G1 adults are both fecund and fertile. Their G2 progeny could live at restrictive temperatures and a small percentage of these reach the adult stage. Thus a minimal rescue schedule provides the opportunity to study animals homozygous for the Phm^P(28) mutation past their normal lethal phase. A study of G2 animals has produced additional observations on the requirements for Phm activity during later (metamorphic) developmental stages. G2 animals show two prominent developmental defects that are ascribed to insufficient Phm activity -- a prevalent deformity of puparia and a developmental block that occurs during or just after head eversion. This suggests that events at or around this critical stage require Phm (and signaling by amidated peptides) for normal progression to form the puparium and to complete adult development (Jiang, 2000).

The poor disc and head eversion may be due to retained attachment of larval mouthparts: without an ability to move posteriorly within the puparium, the animal faces increased confinement and antagonism to the emergence of pupal tissue. If this explanation is correct, the defects involving Phm activity center more on postpupation events than on pupation itself. Secondary hs-Phm transgene inductions increased the percentage of animals successfully completing these developmental transitions from roughly 20% to nearly 100%. Likewise the form of animals receiving additional inductions more closely resembles that of wild-type animals. These observations support the hypothesis that lowered Phm levels are responsible for these 'late' mutant phenotypes (Jiang, 2000).

The phenotypes produced in hypomorphic Phm mutant animals closely resemble those produced by hypomorphic mutations in several of the ecdysone-response genes. In addition, they resemble those of genes implicated in ecdysone/steroid hormone production, including the dre4 and dare genes. Strong similarities between Phm mutant phenotypes and those of ecdysone production/signaling genes are also evident at larval developmental stages. In particular, the 'double mouthhook' phenotype seen in Phm^P[07623] animals (produced by a failure to complete larval molts) is also displayed by certain mutants of the EcR gene, the dare gene, and the developmental mutant cryptocephal (crc). Later in development, EcR, crc, dare, and Phm hypomorphic phenotypes include a failure to pupariate. Also, some Phm hypomorphs, like those of DHR3, survive to pupal stages but die around the time of head eversion with defects in puparial form, body shortening, and gas bubble movements. These defects suggest problems with proper activation of the muscles needed to produce shortening, puparium formation, and disc eversion (Jiang, 2000).

Together these observations strengthen the argument that amidated secretory peptides are required for signaling events that ensure progression through several critical developmental transitions. The inclusion of Phm phenotypes in this common list suggests that amidated secretory peptides are involved in many of the hormonal signaling events that are initially triggered by the steroid hormone ecdysone. In addition, amidated secretory peptides are likely involved in the signaling events that regulate ecdysone production and titers. It will be of interest therefore to test genetic interactions between the peptide and the steroid signaling pathways. Also, to place Phm defects within the framework of known regulatory pathways, it will be useful to measure the expression of RNAs for various steroid hormone response genes in Phm mutant animals. The paradigms established here should be useful for future screens that seek to identify genes needed to produce and to mediate peptide signaling. In general, such information will be useful in assigning functional roles to peptidergic systems in their interactions with steroid hormones and will further define the regulation of insect metamorphosis in molecular detail (Jiang, 2000).

These results provide in vivo evidence that Phm is required for peptide alpha-amidating activity throughout the life of Drosophila. Loss-of-function alleles show that this is true in larvae. Phm^P[07623] animals contain (at best) trace levels of Phm enzyme activity and of Phm protein. Further, when assayed using the expression of FMRFamide neuropeptides, Phm mutant animals display little if any staining for amidated neuropeptides, although staining for nonamidated peptides, for nonpeptide transmitters, and for neuropeptide processing enzymes appears normal. From the analysis of Phm mutant animals that were maintained beyond their normal lethal phase to reach pupal and adult ages, the same conclusion is drawn for late developmental stages as well. Limiting the induction of transgenic Phm to just the first larval days fully rescues Phm mutant lethality. However, such rescued adults are still abnormal, i.e., they contain ~20% of normal Phm levels and display abnormal cellular profiles of amidated FMRF peptides. The abnormal cellular pattern is highly reproducible because identified neurons (e.g., OL2 and MP2) lack staining, while other identified neurons (e.g., SP1) stain normally in all animals examined. It is speculated that these patterns reflect similar abnormalities in other amidated neuropeptides, expressed by other sets of neurons. It is concluded that a compensating activity does not appear later in development (at least not in the case of adult CNS neurons) and that the Phm gene represents the principle source of Phm enzyme activity at all developmental stages (Jiang, 2000).

Based on an immunological survey of three amidated peptide systems, it is inferred that Phm^P[07623] mutant animals lack most amidated peptides and therefore lack most functional neuropeptides. In that regard, their locomotor and feeding behaviors appear remarkably normal for the first hours after larval hatching. Many die, associated with a failure to thrive, and their decline probably reflects a loss of function in several systems. There may be a loss of neural drive that is normally modulated by neuropeptides. Also, the death of mutant animals may reflect the lack of organized digestive functions, since Phm and amidated peptides are abundant in midgut epithelia and are likely required for normal gut physiology. The mutant phenotype may reflect an imbalance in the maintenance or use of energy stores by factors such as Adipokinetic hormone or an absence of sufficient hemolymph regulation by cardioacceleratory peptide. The Phm mutant animals currently available do not allow for the destinction between these or other plausible explanations. However, the viability of the strong Phm hypomorphs, their effective lack of amidated peptide stores, and the availability of methods to create Phm mosaics will help to define specific roles for particular peptidergic systems (Jiang, 2000).

GENE STRUCTURE

Southern blot analysis, using Phm cDNA as probe, indicates that Phm sequences are present in single copy in the haploid Drosophila genome. Although precise definition of the Phm start site is lacking, the gene contains at least eight exons. Comparison of the exon/intron structure of Phm to that of the PHM domain of the rat PAM gene indicates a highly similar organization. Six of the seven introns in Phm occur within identical amino acids to those containing introns in mammalian PAM. A further similarity is found on examination of the intron junctions: exons are interrupted by introns that occur between codons (type 0) or within them (type 1, after one nucleotide; type 2, after two nucleotides). In the case of Phm, the rat and Drosophila genes have comparable 'types' at all six conserved intron junctions. Unlike the rat PHM sequences, which map over more than 76 kb, sequences encoding the Phm open reading frame extend over only ~3 kb of genomic DNA. No evidence was found for the occurrence of alternative splicing in the Phm gene. Also, standard low stringency screening methods failed to detect the presence of PAL-encoding sequences near (i.e., within ~10 kb) the identified PHM gene. Thus, the Drosophila Phm gene seems highly homologous to the mammalian PAM gene, except for the fact that the Drosophila protein lacks a PAL-encoding domain (Kolhekar, 1997a).

PROTEIN STRUCTURE

Amino Acids - 365

Structural Domains

The deduced Phm protein begins with a hydrophobic NH2 terminus that has the properties of a signal sequence and displays a great deal of similarity to vertebrate Phms. Over the catalytic core of the enzyme, there is 40% sequence identity between rat and Drosophila Phm and complete identity extending up to nine consecutive residues. Other notable features of similarity include (1) eight cysteine residues found in homologous positions in all PHMs and DBMs, and (2) two histidine-rich sequence clusters (HHM, aa 95-97, and HTH, aa 241-243) that are thought to be important for the binding of copper to the mono-oxygenase (Eipper, 1995). The presence of several other highly conserved regions suggests that these regions play a previously unrecognized role in catalysis. A stop codon occurs immediately past what is recognized as the catalytic core of rat PHM and is followed by a 201 nucleotide 3-untranslated region and a poly(A)+ tail of 27 nucleotides. Although clearly encoding a PHM protein, the Drosophila cDNA encodes a monofunctional enzyme. No cDNAs encoding bifunctional PAM proteins were identified (Kolhekar, 1997).

date revised: 15 April 2001

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.